CN111583162B - Image Enhancement Method Based on Histogram Equalization - Google Patents
Image Enhancement Method Based on Histogram Equalization Download PDFInfo
- Publication number
- CN111583162B CN111583162B CN202010378155.4A CN202010378155A CN111583162B CN 111583162 B CN111583162 B CN 111583162B CN 202010378155 A CN202010378155 A CN 202010378155A CN 111583162 B CN111583162 B CN 111583162B
- Authority
- CN
- China
- Prior art keywords
- histogram
- image
- global
- sub
- level
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 90
- 230000001186 cumulative effect Effects 0.000 claims description 43
- 230000002457 bidirectional effect Effects 0.000 claims description 24
- 238000000926 separation method Methods 0.000 claims description 15
- 230000004927 fusion Effects 0.000 claims description 11
- 240000004282 Grewia occidentalis Species 0.000 claims description 8
- 238000013507 mapping Methods 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 8
- 238000009826 distribution Methods 0.000 claims description 7
- 238000003709 image segmentation Methods 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 230000003321 amplification Effects 0.000 description 4
- 238000003199 nucleic acid amplification method Methods 0.000 description 4
- 230000000903 blocking effect Effects 0.000 description 3
- 238000003672 processing method Methods 0.000 description 3
- 230000002708 enhancing effect Effects 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
Classifications
-
- 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
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20112—Image segmentation details
- G06T2207/20132—Image cropping
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
- Facsimile Image Signal Circuits (AREA)
Abstract
The invention provides an image enhancement method based on histogram equalization, which comprises the following steps: an image input unit inputs an image; the global histogram equalization unit performs global equalization processing on the image and obtains a first image; the local histogram equalization unit performs local equalization processing on the image and obtains a second image; the scene similarity unit calculates the weight occupied by the first image and the weight occupied by the second image; and the image output unit fuses the first image and the second image according to the weight occupied by the first image and the weight occupied by the second image so as to obtain an enhanced image. In the invention, the input graph is processed, the contrast of the image is enhanced, the details of the image can be kept, and the image has good overall brightness.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to an image enhancement method based on histogram equalization.
Background
The histogram of the image is an image which is based on a statistical concept and is used for counting the size and the number of each pixel level in the image, and the distribution condition of the pixel levels of the image can be well shown. The histogram of an image is often applied in image processing methods, especially in image enhancement methods. Histogram equalization is a very representative method among image contrast enhancement methods, and aims to convert relatively concentrated brightness regions in an image histogram into uniform distribution in a larger range, so that the contrast of an image is effectively enhanced. Because of the advantages of simple histogram equalization principle, low calculation complexity and the like, the method is widely applied to the fields of high-dynamic images, infrared images, underwater image enhancement methods and the like.
The conventional histogram equalization method is classified into a global histogram equalization method and a local histogram equalization method. Global histogram equalization aims at mapping all pixels at the same pixel level in the whole image to another pixel level, which is to contrast-enhance the image in the global scope, but in some cases, this process has problems such as oversaturation of the region, insufficient contrast of the image, and failure to preserve the details of the image. The local histogram equalization is to divide the original image into sub-image blocks and perform histogram equalization processing on the sub-image blocks, so that the processing can effectively improve the definition, and the contrast enhancement effect is better than the global histogram equalization, but the problems of blocking effect, amplified noise in dark areas and the like sometimes exist.
Disclosure of Invention
The invention aims to provide an image enhancement method based on histogram equalization, which is used for processing an input image, enhancing the contrast of the image, keeping the details of the image and providing the overall brightness of the image.
In order to achieve the above object, the present invention provides an image enhancement method based on histogram equalization, the image enhancement method comprising the steps of:
An image input unit inputs an image;
the global histogram equalization unit performs global equalization processing on the image and obtains a first image;
the local histogram equalization unit performs local equalization processing on the image and obtains a second image;
the scene similarity unit calculates the weight occupied by the first image and the weight occupied by the second image;
and the image output unit fuses the first image and the second image according to the weight occupied by the first image and the weight occupied by the second image so as to obtain an enhanced image.
Optionally, in the image enhancement method based on histogram equalization, the method for global equalization processing of the image by the global histogram equalization unit to obtain the first image includes:
the global histogram calculation module counts the pixel distribution of the whole image and forms a global histogram;
the histogram segmentation module segments the global histogram into a global high-level sub-histogram and a global low-level sub-histogram;
the bidirectional limited calculation module calculates the dynamic range of limiting the global high-level sub-histogram and the dynamic range of limiting the global low-level sub-histogram respectively;
and the bidirectional limited equalization module equalizes the global high-level sub-histogram according to the dynamic range of the global high-level sub-histogram, equalizes the global low-level sub-histogram according to the dynamic range of the global low-level sub-histogram, fuses the equalized global high-level sub-histogram and the equalized global low-level sub-histogram, and then applies the fused global high-level sub-histogram and the equalized global low-level sub-histogram to the input image to obtain a first image.
Optionally, in the image enhancement method based on histogram equalization, the method for forming a global histogram by counting the pixel distribution of the whole image by the global histogram calculation module includes:
calculating the total number of pixel points of the whole image;
counting the number of pixels corresponding to the pixel grade according to the pixel grade of the image;
and forming a global histogram by the pixel grade and the ratio of the number of pixels corresponding to the pixel grade to the total number of the pixel points.
Optionally, in the image enhancement method based on histogram equalization, the method for dividing the global histogram into a global high-level sub-histogram and a global low-level sub-histogram by a histogram dividing module includes:
calculating a threshold value of a separation point of the global high-level sub-histogram and the global low-level sub-histogram according to the pixel level and the pixel level occupation ratio of the global histogram;
the threshold value serves as a reference point separating the global histogram into a global high-level sub-histogram and a global low-level sub-histogram.
Optionally, in the histogram equalization-based image enhancement method, the method for respectively calculating the dynamic range of the global high-level sub-histogram and the dynamic range of the global low-level sub-histogram by the bidirectional limited calculation module includes:
And utilizing the threshold value, the highest pixel level of the image, the lowest pixel level of the image and the number of pixel levels of the coding format corresponding to the image to obtain the dynamic range of limiting the global high-level sub-histogram and the dynamic range of limiting the global low-level sub-histogram.
Optionally, in the histogram equalization-based image enhancement method, a formula for limiting the dynamic range of the global high-order sub-histogram to be satisfied is calculated as follows:
wherein: alpha H1 、α H2 For a bi-directional limited intensity parameter configured by a user, the input range is [0,3]The method comprises the steps of carrying out a first treatment on the surface of the n is the number of pixel levels of the coding format corresponding to the image; th is the threshold of the separation point; histMin H Limiting the lower limit of the dynamic range of the global high sub-histogram; histMax (Max) H Limiting the upper limit of the dynamic range of the global high-level sub-histogram; histMin is the lowest pixel level of the image; histMax is the highest pixel level of an image.
Optionally, in the histogram equalization-based image enhancement method, a formula for limiting the dynamic range of the global low-level sub-histogram to be satisfied is calculated as follows:
wherein: alpha L1 、α L2 For a bi-directional limited intensity parameter configured by a user, the input range is [0,3]The method comprises the steps of carrying out a first treatment on the surface of the n is the number of pixel levels of the coding format corresponding to the image; th is the threshold of the separation point; histMin L To limit the lower limit of the dynamic range of the global low-level sub-histogram; histMax (Max) L To limit the upper limit of the dynamic range of the global low-level sub-histogram; histMin is the lowest pixel level of the image; histMax is the highest pixel level of an image.
Optionally, in the histogram equalization-based image enhancement method, the method for equalizing the global high-level sub-histogram by the bidirectional limited equalization module according to the dynamic range of the global high-level sub-histogram and equalizing the global low-level sub-histogram according to the dynamic range of the global low-level sub-histogram includes:
expanding the global high-level sub-histogram according to the dynamic range of the global high-level sub-histogram, and expanding the global low-level sub-histogram according to the dynamic range of the global low-level sub-histogram;
counting the corresponding occupancy value of each pixel level in the global low-level sub-histogram, clipping the occupancy value which is larger than a first preset value, and adding the accumulated clipping value to each occupancy value in the extended global low-level sub-histogram on average;
and counting the occupancy value corresponding to each pixel level in the global high-level sub-histogram, clipping the occupancy value which is larger than a second preset value, and adding the accumulated clipping value to each occupancy value in the extended global high-level sub-histogram.
Optionally, in the histogram equalization-based image enhancement method, the method for fusing the equalized global high-level sub-histogram and the equalized global low-level sub-histogram and then applying the fused global high-level sub-histogram and the equalized global low-level sub-histogram to the input image to obtain the first image includes:
obtaining a cumulative histogram of the extended global high-level sub-histogram and a cumulative histogram of the extended global low-level sub-histogram;
fusing the cumulative histogram of the extended global high-level sub-histogram and the cumulative histogram of the extended global low-level sub-histogram;
and mapping the integrated cumulative histogram as a mapping curve to the input image to obtain a first image.
Optionally, in the image enhancement method based on histogram equalization, the method for obtaining the second image by performing local equalization processing on the image by the local histogram equalization unit includes:
the local image segmentation module equally divides the image into a plurality of sub-images;
the local histogram calculation module counts the pixels of each sub-image;
the local histogram segmentation module divides the sub-image into a local high-order sub-image and a local low-order sub-image;
the local bidirectional limited calculation module calculates the dynamic range of limiting the local high-order sub-image and the dynamic range of limiting the local low-order sub-image respectively;
The local bidirectional limited equalization module equalizes the local high-order sub-image according to the dynamic range of the local high-order sub-image, equalizes the local low-order sub-image according to the dynamic range of the local low-order sub-image, fuses the equalized local high-order sub-histogram with the equalized local low-order sub-histogram, and applies the fused local high-order sub-histogram to the sub-image to obtain a sub-image after local equalization;
and the local image fusion module carries out linear interpolation processing on the sub-images subjected to the local equalization processing and fuses the sub-images subjected to the linear interpolation processing.
Optionally, in the histogram equalization-based image enhancement method, the method for performing linear interpolation processing on the sub-image after the local equalization processing by the local image fusion module includes:
dividing the sub-image subjected to the local equalization treatment into an edge area, a four-corner area and a middle area;
and respectively carrying out interpolation operation on the pixel points of the edge area, the pixel points of the four corner areas and the pixel points of the middle area.
Optionally, in the histogram equalization-based image enhancement method, the method for fusing the first image and the second image by the scene similarity unit includes:
The scene similarity calculation module acquires an information window which takes the current pixel point as a center according to window parameters configured by a user, and calculates a scene similarity value;
the similarity weight calculation module judges whether the current pixel is in a uniform pixel area according to the value of the scene similarity;
if the global histogram is in the uniform pixel region, the global histogram has a larger weight; if in the non-uniform pixel area, the local histogram is weighted more heavily.
Optionally, in the histogram equalization-based image enhancement method, the method for obtaining the value of the scene similarity includes:
calculating absolute differences of pixel grades of all pixel points in the information window and the current pixel point;
and acquiring the weight of the first image and the weight of the second image according to the absolute difference value.
Optionally, in the histogram equalization-based image enhancement method, the weight of the first image is proportional to the value of the scene similarity.
In the image enhancement method based on histogram equalization, global histogram equalization processing is carried out on an input image, meanwhile, local histogram equalization processing is carried out on the input image, and then the image subjected to the global histogram equalization processing and the image subjected to the local histogram equalization processing are combined according to the value of scene similarity to output an enhanced image, so that the contrast of the image is enhanced; the image detail can be kept; and at the same time, the image has good overall brightness.
Furthermore, in the local histogram equalization processing of the invention, the local high-order sub-image and the local low-order sub-image are respectively subjected to bidirectional limitation so as to equalize the local high-order sub-image and the local low-order sub-image, and the problems of supersaturation, noise amplification in a dark area and the like can be suppressed, so that the overall effect of the image is further improved.
Drawings
FIG. 1 is a flow chart of a histogram equalization-based image enhancement method of an embodiment of the present invention;
FIGS. 2 through 4 are process diagrams of histogram variation for a method of enhancing image contrast for histogram equalization in accordance with an embodiment of the present invention;
FIG. 5 is a cumulative histogram map curve of the first image;
FIG. 6 is a schematic diagram of a 5x5 information window with D12 as the current pixel according to an embodiment of the present invention;
in the figure: 1-first curve, 2-second curve, 3-third curve, 4-fourth curve, 5-fifth curve, 6-sixth curve, 7-seventh curve, 8-eighth curve.
Detailed Description
Specific embodiments of the present invention will be described in more detail below with reference to the drawings. The advantages and features of the present invention will become more apparent from the following description. It should be noted that the drawings are in a very simplified form and are all to a non-precise scale, merely for convenience and clarity in aiding in the description of embodiments of the invention.
In the following, the terms "first," "second," and the like are used to distinguish between similar elements and are not necessarily used to describe a particular order or chronological order. It is to be understood that such terms so used are interchangeable under appropriate circumstances. Similarly, if a method described herein comprises a series of steps, and the order of the steps presented herein is not necessarily the only order in which the steps may be performed, and some of the described steps may be omitted and/or some other steps not described herein may be added to the method.
Referring to fig. 1, the present invention provides an image enhancement method based on histogram equalization, comprising:
s11: an image input unit inputs an image;
s12: the global histogram equalization unit performs global equalization on the image to obtain a first image;
s13: the local histogram equalization unit performs local equalization processing on the image to obtain a second image;
s14: the scene similarity unit calculates the weight occupied by the first image and the weight occupied by the second image;
s15: and the image output unit fuses the first image and the second image according to the weight occupied by the first image so as to obtain an enhanced image.
The first image refers to a globally balanced image, and the second image refers to a globally balanced image.
In order to inhibit the problems of supersaturation, noise amplification in dark areas and the like and further improve the overall effect of an image, the invention provides an image enhancement method based on histogram equalization, which can combine the results of global histogram and local histogram equalization processing according to scene similarity information, enhance the contrast of the image, keep the image details and simultaneously enable the image to have good overall brightness, inhibit the problems of supersaturation, noise amplification in dark areas and the like and further improve the overall effect of the image.
Further, referring to fig. 2 to 5, the method for global equalization processing of the image by the global equalization unit to obtain the first image includes:
the global histogram calculation module counts the pixel distribution of the whole image and forms a global histogram;
the histogram segmentation module segments the global histogram into a global high-level sub-histogram and a global low-level sub-histogram;
the bidirectional limited calculation module calculates the dynamic range of limiting the global high-level sub-histogram and the dynamic range of limiting the global low-level sub-histogram respectively;
And the bidirectional limited equalization module equalizes the global high-level sub-histogram according to the dynamic range of the global high-level sub-histogram, equalizes the global low-level sub-histogram according to the dynamic range of the global low-level sub-histogram, and fuses the equalized global high-level sub-histogram and the equalized global low-level sub-histogram to obtain a first image.
Further, the method for forming the global histogram by the global histogram calculation module by counting the pixel distribution of the whole image comprises the following steps:
calculating the total number of pixel points of the whole image;
counting the number of pixels corresponding to the pixel grade according to the pixel grade of the image;
and forming a global histogram by taking the pixel grade as the ratio of the number of pixels corresponding to the pixel grade to the total number of the pixel points.
Specifically, the input image has pixel levels, for example, an 8bit format image is input, the pixel level may be any one of 0 to 255, the same pixel level may have a plurality of pixels, and of course, a situation that one or some pixel levels have no pixels may occur. Each pixel level occupies a certain proportion, and a histogram can be formed by the ratio of the pixel level to the pixel level, and the histograms can be formed by the trend of the histograms, for example, a histogram curve 1 of the input image in fig. 2, the abscissa is the pixel level, and the ordinate is the ratio of the pixel level (the ratio of the number of the pixel levels to the number of the pixel levels of the whole image). Since the histogram curve 1 of the input image has a single peak or multiple peaks, it may affect the equalization of the global histogram, so that the global histogram needs to be equalized, and the first step of the processing is to divide the global histogram into a global high sub-histogram and a global low sub-histogram, and divide the whole histogram into two global sub-histograms with a proper pixel level as a boundary.
Further, the method for dividing the global histogram into the global high-level sub-histogram and the global low-level sub-histogram by the histogram dividing module comprises the following steps:
calculating a threshold value of a separation point of the global high-level sub-histogram and the global low-level sub-histogram according to the pixel level and the pixel level occupation ratio of the global histogram;
the threshold value serves as a reference point separating the global histogram into a global high-level sub-histogram and a global low-level sub-histogram.
Specifically, the threshold value for obtaining the separation point satisfies the following formula:
wherein: the round function here returns a rounded integer value; x is the pixel level of the image; pdf_ori (x) is the fraction of the pixel level of the image; m is the highest pixel level of the coding format to which the image corresponds.
For example, the embodiment of the invention inputs an 8-bit image, the pixel level of the 8-bit image should be a value of 0-255, and the highest pixel level is 255, that is, the value of m is 255, and then 256 levels are all used. The values of the pixel levels of the image input by the embodiment of the invention are all within 0-255, but not every level in 0-255 is necessarily, for example, there are no pixel points with the pixel level of 0 and the pixel level of 1, and there are pixel points with the pixel level of 3 and the pixel level of 4. And obtaining the product of the occupation ratio of the pixel level 3 and the pixel level 3, obtaining the product of the occupation ratio of the pixel level 4 and the pixel level 4, and summing the products of the pixel level 3 and the pixel level 4 to form an integer which is the threshold value of the separation point. If in other embodiments of the present invention, an image of other format is input, for example, an image of 64bit format, the pixel points have more pixel levels, the product of each pixel level and the ratio of the pixel levels is sequentially obtained in the same way, and finally all the products are summed, and the integer is taken as the threshold value of the separation point. Then, the maximum pixel level and the minimum pixel level of the image are obtained, as shown in fig. 2, the first curve 1 is a histogram of the input image in the histogram, the minimum value and the maximum value of the histogram correspond to each other on the abscissa, the histogram where the curve from the minimum pixel level to the threshold value of the separation point is located is taken as a global low-level sub-histogram, and the histogram where the curve from the threshold value of the separation point to the maximum pixel level is located is taken as a global high-level sub-histogram. The abscissa of the histogram covers all the levels from 0 to 255, but the abscissa corresponding to the curve does not necessarily cover all the levels, and therefore the minimum pixel level of the image, that is, the minimum value of the abscissa corresponding to the curve, is not the minimum value of the abscissa of the histogram, and similarly, the maximum pixel level of the image is also the minimum value of the abscissa of the histogram. As shown in fig. 3, the second curve 2 is a curve of the global low-level sub-histogram, and the third curve 3 is a curve of the global high-level sub-histogram.
In order to prevent the problems of excessive brightness/darkness and abnormal enhancement of noise in a dark area of an image, the dynamic range of the segmented global high-level sub-histogram and the segmented global low-level sub-histogram after transformation is limited, namely, bidirectional limited equalization of the global high-level sub-histogram and the global low-level sub-histogram is carried out, and the overlapped areas of the global high-level sub-histogram and the global low-level sub-histogram after the bidirectional limited equalization are combined according to a linear weighting method.
Further, the method for respectively calculating and limiting the dynamic range of the global high-level sub-histogram and the dynamic range of the global low-level sub-histogram by the bidirectional limited calculation module comprises the following steps:
and respectively obtaining the dynamic range of the global high-level sub-histogram and the dynamic range of the global low-level sub-histogram by using the threshold value of the separation point, the highest pixel level of the image, the lowest pixel level of the image and the pixel level number of the coding format corresponding to the image.
Specifically, the formula for calculating and limiting the dynamic range of the global high-order sub-histogram to be satisfied is as follows:
wherein: alpha H1 、α H2 For a bi-directional limited intensity parameter configured by a user, the input range is [0,3]The method comprises the steps of carrying out a first treatment on the surface of the n is the number of pixel levels of the coding format corresponding to the image, for example, the value of n in the utility model can be 256, n-1 is 255, and 0-255 is the value range of the pixel level in the embodiment of the utility model; th is the threshold of the separation point; histMin H Limiting the lower limit of the dynamic range of the global high sub-histogram; histMax (Max) H Dynamic limiting global high-level sub-histogramsThe upper limit of the range; histMin is the lowest pixel level of the image; histMax is the highest pixel level of an image.
The formula for computing the dynamic range satisfaction limiting the global low-level sub-histogram is as follows:
wherein: alpha L1 、α L2 For a bi-directional limited intensity parameter configured by a user, the input range is [0,3]The specific value is determined by the user; n is the number of pixel levels of the coding format corresponding to the image, for example, the value of n in the utility model can be 256, n-1 is 255, and 0-255 is the value range of the pixel level in the embodiment of the utility model; th is the threshold of the separation point; histMin L To limit the lower limit of the dynamic range of the global low-level sub-histogram; histMax (Max) L To limit the upper limit of the dynamic range of the global low-level sub-histogram; histMin is the lowest pixel level of the image; histMax is the highest pixel level of an image.
The range of the lower limit of the dynamic range of the global low-level sub-histogram, the upper limit of the dynamic range of the global low-level sub-histogram, the lower limit of the dynamic range of the global high-level sub-histogram and the upper limit of the dynamic range of the global high-level sub-histogram is between 0 and the maximum pixel level of the coding format corresponding to the image, namely, the range is limited by [0,255] in the embodiment of the utility model. That is, the dynamic range calculated is 255 if it exceeds 255, and 0 if it falls below 0. When the bidirectional limited intensity parameter input by the user is 0, the dynamic range is not expanded, and the larger the bidirectional limited intensity parameter input by the user is, the more the dynamic range is expanded, the larger the brightness contrast of the image is enhanced.
Further, the method for equalizing the global high-level sub-histogram by the bidirectional limited equalization module according to the dynamic range of the global high-level sub-histogram and equalizing the global low-level sub-histogram according to the dynamic range of the global low-level sub-histogram comprises the following steps:
expanding the global high-level sub-histogram according to the dynamic range of the global high-level sub-histogram, and expanding the global low-level sub-histogram according to the dynamic range of the global low-level sub-histogram;
counting the occupation ratio corresponding to each pixel level in the [ histMin, th ] of the global low-level sub-histogram, clipping the occupation ratio which is larger than a first preset value, and adding the accumulated clipping value to each occupation ratio in the [ histMinL, histMaxL ] of the extended global low-level sub-histogram on average;
and counting the occupation ratio corresponding to each pixel level in the global high-level sub-histogram [ Th, histMax ], clipping the occupation ratio larger than a second preset value, and adding the accumulated clipping value to each occupation ratio in the extended global high-level sub-histogram [ histMinH, histMaxH ] on average. The lower limit of the dynamic range of the global high-level sub-histogram and the upper limit of the dynamic range of the global high-level sub-histogram are taken as the minimum level and the maximum level of the curve in the expanded global high-level sub-histogram after the equalization processing, and the lower limit of the dynamic range of the global low-level sub-histogram and the upper limit of the dynamic range of the global low-level sub-histogram are taken as the minimum level and the maximum level of the curve in the expanded global low-level sub-histogram after the equalization processing. The first preset value is a clipping value configured by a user, defaulting to 0.01, if the global low-level sub-histogram has a value exceeding 0.01 in the ratio value corresponding to each pixel level in [ histMin, th ], clipping the global low-level sub-histogram, for example, if the value corresponding to a certain pixel level is 0.015, the clipped value is 0.005, that is, the value obtained by subtracting 0.01 from 0.015. If there are multiple levels of corresponding duty ratios to be clipped, the sum of the clipped values is added to the duty ratio value corresponding to each pixel level in the extended global low-level sub-histogram [ histMinL, histMaxL ]. As shown in fig. 4, the fourth curve 4 is the equalized global low-level sub-histogram, and the fifth curve 5 is the equalized global high-level sub-histogram.
Further, the method for fusing the equalized global high-level sub-histogram and the equalized global low-level sub-histogram includes:
respectively obtaining a cumulative histogram of the extended global high sub-histogram and a cumulative histogram of the extended global low sub-histogram;
the cumulative histogram of the extended global high-level sub-histogram and the cumulative histogram of the extended global low-level sub-histogram are fused.
Specifically, the expression of the cumulative histogram of the extended global high-order sub-histogram is found as follows:
wherein: x represents the pixel level of the histogram, the range is [0,255], cdf_h (x) is the pixel occupation value of the cumulative histogram of the extended global high-level sub-histogram, pdf_h (i) is the value of the pixel level of the high-level sub-histogram, that is, the interpretation of this formula is that when i takes a certain pixel level, the sum of the pixel levels of all the corresponding high-level sub-histograms starting from 0 to i is taken as the pixel occupation value of the cumulative histogram of the extended global high-level sub-histogram, and as the value of i changes, the value of cdf_h (x) also changes continuously, so as to form the cumulative histogram of the global high-level sub-histogram.
The cumulative histogram of the extended global low-level sub-histogram is expressed as follows:
Wherein: x represents the pixel level of the histogram, the range is [0,255], cdf_l (x) is the pixel occupation value of the cumulative histogram of the extended global low-level sub-histogram, pdf_l (i) is the value of the pixel level of the low-level sub-histogram, that is, the interpretation of this formula is that when i takes a certain pixel level, the sum of the pixel levels of all the corresponding low-level sub-histograms starting from 0 to i is taken as the pixel occupation value of the cumulative histogram of the extended global high-level sub-histogram, and as the value of i changes, the value of cdf_l (x) also changes continuously, so as to form the cumulative histogram of the global low-level sub-histogram.
Then according to the expression, when x is less than histMin L The expression for obtaining the integrated cumulative histogram is as follows:
cdf_M(x)＝x，
wherein: x is the pixel level; cdf_m (x) is the cumulative histogram after fusion;
while the histMin L ≤x＜histMin H The expression for obtaining the integrated cumulative histogram is as follows:
cdf_M(x)＝cdf_L(x)，
wherein: cdf_l (x) is the pixel occupancy value of the cumulative histogram of the extended global low-level sub-histogram; cdf_m (x) is the cumulative histogram after fusion;
while the histMin H ≤x＜histMax L The expression for obtaining the integrated cumulative histogram is as follows:
wherein: cdf_l (x) is the pixel occupancy value of the cumulative histogram of the extended global low-level sub-histogram; histMin H Is the lower limit of the dynamic range of the global high-level sub-histogram; histMax (Max) L To limit the upper limit of the dynamic range of the global low-level sub-histogram; histMin H Is the lower limit of the dynamic range of the global high-level sub-histogram; histMax (Max) L Is the upper limit of the dynamic range of the global low-level sub-histogram; cdf_h (x) is the pixel occupancy value of the cumulative histogram of the extended global high-order sub-histogram; cdf_m (x) is the cumulative histogram after fusion;
when histMax L ≤x＜histMax H The expression for obtaining the integrated cumulative histogram is as follows:
cdf_M(x)＝cdf_H(x)，
wherein: cdf_h (x) is the pixel occupancy value of the cumulative histogram of the extended global high-order sub-histogram; cdf_m (x) is the cumulative histogram after fusion;
when x is greater than or equal to histMax H The expression for obtaining the integrated cumulative histogram is as follows:
cdf_M(x)＝x，
wherein: x is the pixel level; cdf_m (x) is the cumulative histogram after fusion.
The fused image is divided into a plurality of segments, the mapping expression of each segment is different, the peak is eliminated, the curve of the histogram of the image becomes more uniform, and the fused cumulative histogram cdf_M (x) is used as a mapping curve to be mapped onto the input image, so that a first image is obtained. Fig. 5 is a cumulative histogram map curve of the first image, the abscissa is the pixel level, the ordinate is the mapped pixel level of the cumulative histogram, the sixth curve 6 is the mapped curve of the cumulative histogram of the extended global low sub-histogram, the seventh curve 7 is the mapped curve of the cumulative histogram of the extended global high sub-histogram, and the eighth curve 8 is the mapped curve of the fused cumulative histogram.
Further, the method for obtaining the second image by performing local equalization processing on the image by the local histogram equalization unit includes:
the local image segmentation module equally divides the image into a plurality of sub-images;
the local histogram calculation module counts pixels of the sub-image;
the local histogram segmentation module divides the sub-image into a local high-order sub-image and a local low-order sub-image;
the local bidirectional limited calculation module calculates and limits the dynamic range of the local high-order sub-image and the dynamic range of the local low-order sub-image respectively;
the local bidirectional limited equalization module equalizes the local high-order sub-image according to the dynamic range of the local high-order sub-image, equalizes the local low-order sub-image according to the dynamic range of the local low-order sub-image, fuses the equalized local high-order sub-histogram and the equalized local low-order sub-histogram, and applies the fused local high-order sub-histogram and the equalized local low-order sub-histogram to the sub-image to obtain a sub-image after local equalization;
and the local image fusion module carries out linear interpolation processing on the sub-images subjected to the local equalization processing and fuses the sub-images subjected to the linear interpolation processing.
Since the image contrast is globally enhanced during the global histogram equalization process, some local features are not desirably enhanced. In order to further enhance the details of the image and at the same time enhance the local contrast of the image, it is necessary to further perform local histogram equalization processing on the input image. The local equalization processing method needs to divide an input image into a plurality of local modules and then perform equalization processing on each local module separately. The length and width can be divided into 8 parts according to the length and width values of the input image, and 8×8=64 partial images which maintain the aspect ratio of the input image can be obtained in total from 8 rows and 8 columns. The local histogram calculation module counts the pixels of each sub-image; the local histogram segmentation module divides the sub-image into a local high-order sub-image and a local low-order sub-image; the local bidirectional limited calculation module calculates the dynamic ranges of the limited local high-level sub-image and the limited local low-level sub-image respectively; the local bidirectional limited equalization module equalizes the local high-order sub-image according to the dynamic range of the local high-order sub-image and equalizes the local low-order sub-image according to the dynamic range of the local low-order sub-image in the same way as the global histogram equalization method, and therefore, the detailed description thereof is omitted.
Further, the method for linear interpolation processing of the sub-images after the local equalization processing by the local image fusion module comprises the following steps: dividing the sub-image subjected to the local equalization treatment into an edge area, a four-corner area and a middle area; and respectively carrying out interpolation operation on the pixel points of the edge area, the pixel points of the four corner areas and the pixel points of the middle area.
Since the local equalization processing method processes the individual local images, a very obvious blocking effect remains between the sub-images, so that linear interpolation processing between the sub-images is required, so that the value after each pixel point transformation is obtained by bilinear interpolation of 4 sub-images around the value to eliminate the influence of the blocking effect. The pixels in the four-corner areas refer to the pixels located at four corners in the pixel block diagram of the sub-image, and for the pixels in the four-corner areas, the calculated pixels are still the pixels. The pixel points of the edge area refer to the pixel points of the edge in the pixel block diagram of the sub-image, and the calculated value is the interpolation of the pixel points and the adjacent edge pixels of the adjacent partial image for the pixel points of the edge area. The pixel points of the middle region refer to the pixel points located in the middle of the pixel block diagram of the sub-image. And for the pixel points in the middle area, the calculated pixel points are the bilinear interpolation of the mapping curves of the current sub-image and the three sub-images which are nearest to the current sub-image. After the pixel point of each sub-image is processed, the sub-images after the linear interpolation processing are fused together, and a second image is obtained.
Further, the method for fusing the first image and the second image by the scene similarity unit includes:
the scene similarity calculation module acquires an information window with the current pixel point as a center according to window parameters configured by a user, and calculates a value of scene similarity;
the similarity weight calculation module judges whether the current pixel is in a uniform pixel area according to the value of the scene similarity;
the global histogram is weighted more heavily if in the uniform pixel region and the local histogram is weighted more heavily if in the non-uniform pixel region.
Further, the method for obtaining the value of the scene similarity comprises the following steps:
calculating absolute differences between all pixel points in the information window and the current pixel point;
and respectively acquiring the weight of the first image and the weight of the second image according to the absolute difference value.
Specifically, the method for acquiring the information window with the current point as the center according to the window parameters configured by the user comprises the following steps: taking the window parameter wsize=5 as an example, as shown in fig. 6, a 5×5 information window centered on the current pixel D12 calculates a value of the scene similarity according to the pixel level of the current pixel and its neighboring pixels. The formula for calculating the absolute difference value of the pixel level of all the pixel points in the information window and the current pixel point is as follows:
AbsDiff(x)＝|D(x)-D(12)|，
Wherein: absDiff (x) is the absolute difference value of the pixel level of all the pixel points in the information window and the current pixel point; d (x) is the pixel level of the adjacent pixel point of the current pixel point; d (12) is the pixel level of the current pixel point.
Then the value of the scene similarity corresponding to the current pixel point D12 is calculated as follows:
/>
wherein: delta is the global standard deviation of the image, i.e. the standard deviation of the pixel level of the whole image, which can be calculated by the prior art and is not described in detail here; sim is the value of scene similarity; absDiff (x) is the absolute difference of the pixel level of all pixels in the information window from the current pixel.
Further, the weight of the first image is proportional to the value of the scene similarity. If the value of the scene similarity is larger, the current pixel point is in a uniform area, good brightness is expected to be kept, and more data of the image subjected to global histogram equalization processing should be combined, so that the given weight is larger. If the scene similarity value of the current pixel point is smaller, the current pixel point is in a non-uniform area (detail area), and the data of the image after the local histogram equalization processing should be mainly combined, so that the given weight is smaller.
More specifically, the calculation formula for calculating the weight of the first image according to the value of the scene similarity is as follows:
w_sim=p×sim, where p is a weight coefficient configured by the user, and the default value is 0.4; w_sim is the weight of the first image; sim is the value of the scene similarity.
Finally, the image output unit fuses the first image and the second image according to the weight occupied by the first image and the weight occupied by the second image to obtain an enhanced image, and the satisfied formula is as follows:
i_enhancement=w_sim×i_global+ (1-w_sim) ×i_local, wherein: i_enhancement is an enhanced image; w_sim is the weight of the first image; i_global is a first picture, and i_local is a second picture.
In summary, in the image enhancement method based on histogram equalization provided by the embodiment of the invention, global histogram equalization processing is performed on an input image, meanwhile, local histogram equalization processing is performed on the input image, and then the image subjected to global histogram equalization processing and the image subjected to local histogram equalization processing are combined according to a scene similarity value to output an enhanced image, so that the contrast of the image is enhanced; the image detail can be kept; and at the same time, the image has good overall brightness. Furthermore, in the local histogram equalization processing of the invention, the local high-order sub-image and the local low-order sub-image are respectively subjected to bidirectional limitation so as to equalize the local high-order sub-image and the local low-order sub-image, and the problems of supersaturation, noise amplification in a dark area and the like can be suppressed, so that the overall effect of the image is further improved.
The foregoing is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any person skilled in the art will make any equivalent substitution or modification to the technical solution and technical content disclosed in the invention without departing from the scope of the technical solution of the invention, and the technical solution of the invention is not departing from the scope of the invention.
Claims (11)
1. An image enhancement method based on histogram equalization, characterized in that the image enhancement method comprises the following steps:
an image input unit inputs an image;
the global histogram equalization unit performs global equalization processing on the image and obtains a first image;
the local histogram equalization unit performs local equalization processing on the image and obtains a second image;
the scene similarity unit calculates the weight occupied by the first image and the weight occupied by the second image;
the image output unit fuses the first image and the second image according to the weight occupied by the first image and the weight occupied by the second image so as to obtain an enhanced image;
the method for obtaining the first image by carrying out global equalization processing on the image by the global histogram equalization unit comprises the following steps:
The global histogram calculation module counts the pixel distribution of the whole image and forms a global histogram;
the histogram segmentation module segments the global histogram into a global high-level sub-histogram and a global low-level sub-histogram;
the bidirectional limited calculation module calculates the dynamic range of limiting the global high-level sub-histogram and the dynamic range of limiting the global low-level sub-histogram respectively;
the bidirectional limited equalization module equalizes the global high-level sub-histogram according to the dynamic range of the global high-level sub-histogram, equalizes the global low-level sub-histogram according to the dynamic range of the global low-level sub-histogram, fuses the equalized global high-level sub-histogram and the equalized global low-level sub-histogram, and applies the fused global high-level sub-histogram and the equalized global low-level sub-histogram to the input image to obtain a first image;
the local histogram equalization unit performs local equalization processing on the image to obtain a second image, and the method comprises the following steps:
the local image segmentation module equally divides the image into a plurality of sub-images;
the local histogram calculation module counts the pixels of each sub-image;
the local histogram segmentation module divides the sub-image into a local high-order sub-image and a local low-order sub-image;
The local bidirectional limited calculation module calculates the dynamic range of limiting the local high-order sub-image and the dynamic range of limiting the local low-order sub-image respectively;
the local bidirectional limited equalization module equalizes the local high-order sub-image according to the dynamic range of the local high-order sub-image, equalizes the local low-order sub-image according to the dynamic range of the local low-order sub-image, fuses the equalized local high-order sub-histogram with the equalized local low-order sub-histogram, and applies the fused local high-order sub-histogram to the sub-image to obtain a sub-image after local equalization;
the local image fusion module carries out linear interpolation processing on the sub-images subjected to the local equalization processing and fuses the sub-images subjected to the linear interpolation processing;
the method for fusing the first image and the second image by the scene similarity unit comprises the following steps:
the scene similarity calculation module acquires an information window which takes the current pixel point as a center according to window parameters configured by a user, and calculates a scene similarity value;
the similarity weight calculation module judges whether the current pixel is in a uniform pixel area according to the value of the scene similarity;
if the global histogram is in the uniform pixel region, the global histogram has a larger weight; if in the non-uniform pixel area, the local histogram is weighted more heavily.
2. The histogram equalization-based image enhancement method of claim 1, wherein the method of the global histogram calculation module counting the distribution of pixels throughout the image to form a global histogram comprises:
calculating the total number of pixel points of the whole image;
counting the number of pixels corresponding to the pixel grade according to the pixel grade of the image;
and forming a global histogram by the pixel grade and the ratio of the number of pixels corresponding to the pixel grade to the total number of the pixel points.
3. The histogram equalization-based image enhancement method of claim 2, wherein the method of a histogram segmentation module segmenting the global histogram into a global high-level sub-histogram and a global low-level sub-histogram comprises:
calculating a threshold value of a separation point of the global high-level sub-histogram and the global low-level sub-histogram according to the pixel level and the pixel level occupation ratio of the global histogram;
the threshold value serves as a reference point separating the global histogram into a global high-level sub-histogram and a global low-level sub-histogram.
4. The histogram equalization based image enhancement method of claim 3, wherein the method of bi-directionally constrained computing module computing the constraint on the dynamic range of the global high-level sub-histogram and the constraint on the dynamic range of the global low-level sub-histogram, respectively, comprises:
And utilizing the threshold value, the highest pixel level of the image, the lowest pixel level of the image and the number of pixel levels of the coding format corresponding to the image to obtain the dynamic range of limiting the global high-level sub-histogram and the dynamic range of limiting the global low-level sub-histogram.
5. The histogram equalization-based image enhancement method of claim 4, wherein a formula is calculated that limits the dynamic range satisfaction of said global high-order sub-histogram as follows:
wherein: alpha H1 、α H2 For a bi-directional limited intensity parameter configured by a user, the input range is [0,3]The method comprises the steps of carrying out a first treatment on the surface of the n is the number of pixel levels of the coding format corresponding to the image; th is the threshold of the separation point; histMin H Limiting the lower limit of the dynamic range of the global high sub-histogram; histMax (Max) H Limiting the upper limit of the dynamic range of the global high-level sub-histogram; histMin is the lowest pixel level of the image; histMax is the highest pixel level of an image.
6. The histogram equalization-based image enhancement method of claim 5, wherein a formula is calculated that limits the dynamic range satisfaction of the global low-level sub-histogram as follows:
wherein: alpha L1 、α L2 For a bi-directional limited intensity parameter configured by a user, the input range is [0,3 ]The method comprises the steps of carrying out a first treatment on the surface of the n is the number of pixel levels of the coding format corresponding to the image; th is the threshold of the separation point; histMin L To limit the lower limit of the dynamic range of the global low-level sub-histogram; histMax (Max) L To limit the upper limit of the dynamic range of the global low-level sub-histogram; histMin is the lowest pixel level of the image; histMax is the highest pixel level of an image.
7. The histogram equalization-based image enhancement method of claim 6, wherein the method of bi-directionally constrained equalization module equalizing the global high-order sub-histogram according to the dynamic range of the global high-order sub-histogram and equalizing the global low-order sub-histogram according to the dynamic range of the global low-order sub-histogram comprises:
expanding the global high-level sub-histogram according to the dynamic range of the global high-level sub-histogram, and expanding the global low-level sub-histogram according to the dynamic range of the global low-level sub-histogram;
counting the corresponding occupancy value of each pixel level in the global low-level sub-histogram, clipping the occupancy value which is larger than a first preset value, and adding the accumulated clipping value to each occupancy value in the extended global low-level sub-histogram on average;
And counting the occupancy value corresponding to each pixel level in the global high-level sub-histogram, clipping the occupancy value which is larger than a second preset value, and adding the accumulated clipping value to each occupancy value in the extended global high-level sub-histogram.
8. The histogram equalization-based image enhancement method of claim 7, wherein the method of fusing the equalized global high-order sub-histogram and the equalized global low-order sub-histogram and then applying the fused global high-order sub-histogram to the input image to obtain a first image comprises:
obtaining a cumulative histogram of the extended global high-level sub-histogram and a cumulative histogram of the extended global low-level sub-histogram;
fusing the cumulative histogram of the extended global high-level sub-histogram and the cumulative histogram of the extended global low-level sub-histogram;
and mapping the integrated cumulative histogram as a mapping curve to the input image to obtain a first image.
9. The histogram equalization-based image enhancement method of claim 8, wherein the method of performing linear interpolation processing on the sub-image after the local equalization processing by the local image fusion module comprises:
Dividing the sub-image subjected to the local equalization treatment into an edge area, a four-corner area and a middle area;
and respectively carrying out interpolation operation on the pixel points of the edge area, the pixel points of the four corner areas and the pixel points of the middle area.
10. The histogram equalization-based image enhancement method of claim 9, wherein the method of obtaining the value of the scene similarity comprises:
calculating absolute differences of pixel grades of all pixel points in the information window and the current pixel point;
and acquiring the weight of the first image and the weight of the second image according to the absolute difference value.
11. The histogram equalization-based image enhancement method of claim 10, wherein the weight of the first image is proportional to the value of the scene similarity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010378155.4A CN111583162B (en) | 2020-05-07 | 2020-05-07 | Image Enhancement Method Based on Histogram Equalization |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010378155.4A CN111583162B (en) | 2020-05-07 | 2020-05-07 | Image Enhancement Method Based on Histogram Equalization |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111583162A CN111583162A (en) | 2020-08-25 |
| CN111583162B true CN111583162B (en) | 2023-09-01 |
Family
ID=72112049
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010378155.4A Active CN111583162B (en) | 2020-05-07 | 2020-05-07 | Image Enhancement Method Based on Histogram Equalization |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111583162B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105654438A (en) * | 2015-12-27 | 2016-06-08 | 西南技术物理研究所 | Gray scale image fitting enhancement method based on local histogram equalization |
| CN107301635A (en) * | 2017-06-28 | 2017-10-27 | 武汉格物优信科技有限公司 | A kind of infrared image detail enhancing method and device |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8731322B2 (en) * | 2007-05-03 | 2014-05-20 | Mtekvision Co., Ltd. | Image brightness controlling apparatus and method thereof |
-
2020
- 2020-05-07 CN CN202010378155.4A patent/CN111583162B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105654438A (en) * | 2015-12-27 | 2016-06-08 | 西南技术物理研究所 | Gray scale image fitting enhancement method based on local histogram equalization |
| CN107301635A (en) * | 2017-06-28 | 2017-10-27 | 武汉格物优信科技有限公司 | A kind of infrared image detail enhancing method and device |
Non-Patent Citations (1)
| Title |
|---|
| 江巨浪 ; 张佑生 ; 薛峰 ; 胡敏 ; .保持图像亮度的局部直方图均衡算法.电子学报.2006,(05),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111583162A (en) | 2020-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10424054B2 (en) | Low-illumination image processing method and device | |
| US20170365046A1 (en) | Algorithm and device for image processing | |
| Pham et al. | Separable bilateral filtering for fast video preprocessing | |
| Kim et al. | Optimized contrast enhancement for real-time image and video dehazing | |
| JP4423298B2 (en) | Text-like edge enhancement in digital images | |
| JP4858609B2 (en) | Noise reduction device, noise reduction method, and noise reduction program | |
| US9147238B1 (en) | Adaptive histogram-based video contrast enhancement | |
| US8831346B2 (en) | Image processing apparatus and method, and program | |
| CN111915613A (en) | Image instance segmentation method, device, equipment and storage medium | |
| CN107784637B (en) | Infrared image enhancement method | |
| CN101877123A (en) | Image enhancement method and device | |
| EP3203439B1 (en) | Method and device for reducing noise in a component of a picture | |
| US20120321214A1 (en) | Image processing apparatus and method, program, and recording medium | |
| CN115131375B (en) | Automatic ore cutting method | |
| CN111161177B (en) | Image self-adaptive noise reduction method and device | |
| JP2010130150A (en) | Gray level correction apparatus and imaging apparatus | |
| Qi et al. | A neutrosophic filter for high-density salt and pepper noise based on pixel-wise adaptive smoothing parameter | |
| CN105279742B (en) | A kind of image de-noising method quickly based on piecemeal estimation of noise energy | |
| Ramya et al. | A novel method for the contrast enhancement of fog degraded video sequences | |
| CN111583162B (en) | Image Enhancement Method Based on Histogram Equalization | |
| CN113256533B (en) | Self-adaptive low-illumination image enhancement method and system based on MSRCR | |
| KR20190058753A (en) | Image processing method and image processor performing the same | |
| CN103136735A (en) | Single image defogging method based on dual-scale dark channel | |
| Tang et al. | Sky-preserved image dehazing and enhancement for outdoor scenes | |
| CN114066916A (en) | Detection and segmentation method of adhered rice grains based on deep learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |